Ablo Sync Engine HTTP API

The HTTP surface of the Ablo sync engine. Most production traffic flows over WebSocket via @abloatai/ablo; this HTTP API is the Vercel-safe transport for environments that cannot hold long-lived sockets, plus the control plane for capabilities and usage.

The per-model resource routes (/v1/models/{model}) are generated from your own pushed schema — they are documented here generically over {model}, with tasks as the running example. Authenticate every request with a Bearer API key or scoped token.

State

Atomic multi-model commits.

• POSTCommit one or more mutations/v1/commits

Control plane

Capabilities and usage.

• GETList capabilities/v1/capabilities
• POSTMint a capability/v1/capabilities
• GETRetrieve a capability/v1/capabilities/{id}
• DELETERevoke a capability/v1/capabilities/{id}
• GETRead usage/v1/usage

Records

Typed CRUD over any model in your schema.

• GETList records of a model/v1/models/{model}
• POSTCreate a record/v1/models/{model}
• GETRetrieve a record/v1/models/{model}/{id}
• DELETEDelete a record/v1/models/{model}/{id}
• PATCHUpdate a record/v1/models/{model}/{id}

Coordination

Claims — pessimistic row reservation for human+agent coordination.

• POSTClaim a record/v1/models/{model}/{id}/claim
• DELETERelease a claim/v1/models/{model}/{id}/claim
• POSTHeartbeat a claim/v1/models/{model}/{id}/claim/heartbeat
• POSTReorder the claim wait-line (privileged)/v1/models/{model}/{id}/claim/reorder